Merge pull request #55 from fracai/mac_support

Mac support

CHANGELOG.md

@@ -0,0 +1,111 @@
+# Changelog
+All notable changes to this project will be documented in this file.
+
+The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
+and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+This changlog uses the [ISO 8601 date format](https://www.iso.org/iso-8601-date-and-time-format.html) of (YYYY-MM-DD).
+
+## [Unreleased]
+
+## [1.5.1] - 2021-06-05
+
+### Changed
+
+* Moved Changelog from Readme to Changelog.md file
+
+### Fixed
+
+* Fix TypeError when reading playlist files [#50](https://github.com/nims11/IPod-Shuffle-4g/pull/50)
+
+## [1.5.0] - 2020-06-10
+
+### Changed
+
+* Port Script to Python3
+* Mutagen support is now optional
+
+## [1.4.0] - 2016-08-28
+
+### Added
+
+* Added optional `--verbose` output
+* Added files to `extras` folder
+* Added shortcut parameters (`-p`, `-t`, `-d`, etc.)
+
+### Changed
+
+* Renamed `--voiceover` to `--track-voiceover`
+* Renamed script from `shuffle.py` to `ipod-shuffle-4g.py`
+* Ignore hidden filenames
+* Do not force playlist voiceover with auto playlists
+
+### Fixed
+
+* Catch "no space left" error [#30](https://github.com/nims11/IPod-Shuffle-4g/issues/30)
+* Fix UnicodeEncodeError for non-ascii playlist names [#35](https://github.com/nims11/IPod-Shuffle-4g/issues/35)
+
+## [1.3.0] - 2016-06-08
+
+### Added
+
+* Directory based auto playlist building (`--auto-dir-playlists`) [#13](https://github.com/nims11/IPod-Shuffle-4g/issues/13)
+* ID3 tags based auto playlist building (`--auto-id3-playlists`)
+* Added short program description
+* Differentiate track and playlist voiceover [#26](https://github.com/nims11/IPod-Shuffle-4g/issues/26)
+
+### Changed
+
+* Voiceover disabled by default [#26](https://github.com/nims11/IPod-Shuffle-4g/issues/26) (Playlist voiceover enabled with auto playlist generation)
+
+
+### Fixed
+
+* Fix hyphen in filename [#4](https://github.com/nims11/IPod-Shuffle-4g/issues/4)
+* Fixed mutagen bug [#5](https://github.com/nims11/IPod-Shuffle-4g/issues/5)
+
+## [1.2.0] - 2016-02-04
+
+### Added
+
+* Added Path help entry
+* Added MIT License
+* Added this changelog
+
+### Changed
+
+* Skip existing voiceover files with the same name (e.g. "Track 1.mp3")
+* Made help message lower case
+* Improved Readme
+* Improved docs
+
+### Fixed
+
+* Additional fixes from NicoHood
+* Fixed "All Songs" and "Playlist N" sounds when voiceover is disabled [#17](https://github.com/nims11/IPod-Shuffle-4g/issues/17)
+* Better handle broken playlist paths [#16](https://github.com/nims11/IPod-Shuffle-4g/issues/16)
+* Only use voiceover if dependencies are installed
+
+## [1.1.0] - 2016-01-23
+
+### Added
+
+* Fixes from nims11 fork
+* Option to disable voiceover
+* Initialize the IPod Directory tree
+* Using the `--rename-unicode` flag filenames with strange characters and different language are renamed which avoids the script to crash with a Unicode Error
+* Other small fixes
+
+## [1.0.0] - 2012-10-17
+
+### Added
+
+* Original release by ikelos
+
+[Unreleased]: https://github.com/nims11/IPod-Shuffle-4g/compare/1.5.1...HEAD
+[1.5.1]: https://github.com/nims11/IPod-Shuffle-4g/compare/v1.5...1.5.1
+[1.5.0]: https://github.com/nims11/IPod-Shuffle-4g/compare/v1.4...v1.5
+[1.4.0]: https://github.com/nims11/IPod-Shuffle-4g/compare/v1.3...v1.4
+[1.3.0]: https://github.com/nims11/IPod-Shuffle-4g/compare/v1.2...v1.3
+[1.2.0]: https://github.com/nims11/IPod-Shuffle-4g/compare/v1.1...v1.2
+[1.1.0]: https://github.com/nims11/IPod-Shuffle-4g/compare/646b7def4c498c59b063e535a5b64695d8d87e6b...v1.1
+[1.0.0]: https://github.com/nims11/IPod-Shuffle-4g/commit/646b7def4c498c59b063e535a5b64695d8d87e6b

README.md

@@ -1,70 +1,98 @@
-# IPod Shuffle 4g Scripts
+# IPod Shuffle 4g Script
 
-##shuffle.py
+## ipod-shuffle-4g.py
 
 Python script for building the Track and Playlist database for the newer gen IPod Shuffle.
 Forked from the [shuffle-db-ng project](https://code.google.com/p/shuffle-db-ng/)
+and improved my nims11 and NicoHood.
 
-Just put your audio files into the mass storage of your IPod and shuffle.py will do the rest
-```bash
-$ python shuffle.py -h
-usage: shuffle.py [-h] [--disable-voiceover] [--rename-unicode]
-                  [--track-gain TRACK_GAIN]
-                  path
+Just put your audio files into the mass storage of your IPod and shuffle.py will do the rest.
+```
+$ ./ipod-shuffle-4g.py --help
+usage: ipod-shuffle-4g.py [-h] [-t] [-p] [-u] [-g TRACK_GAIN]
+                          [-d [AUTO_DIR_PLAYLISTS]] [-i [ID3_TEMPLATE]] [-v]
+                          path
+
+Python script for building the Track and Playlist database for the newer gen
+IPod Shuffle. Version 1.5
 
 positional arguments:
-  path
+  path                  Path to the IPod's root directory
 
 optional arguments:
   -h, --help            show this help message and exit
-  --disable-voiceover   Disable Voiceover Feature
-  --rename-unicode      Rename Files Causing Unicode Errors, will do minimal
+  -t, --track-voiceover
+                        Enable track voiceover feature
+  -p, --playlist-voiceover
+                        Enable playlist voiceover feature
+  -u, --rename-unicode  Rename files causing unicode errors, will do minimal
                         required renaming
-  --track-gain TRACK_GAIN
-                        Store this volume gain (0-99) for all tracks; 0
-                        (default) means no gain and is usually fine; e.g. 60
-                        is very loud even on minimal player volume
+  -g TRACK_GAIN, --track-gain TRACK_GAIN
+                        Specify volume gain (0-99) for all tracks; 0 (default)
+                        means no gain and is usually fine; e.g. 60 is very
+                        loud even on minimal player volume
+  -d [AUTO_DIR_PLAYLISTS], --auto-dir-playlists [AUTO_DIR_PLAYLISTS]
+                        Generate automatic playlists for each folder
+                        recursively inside "IPod_Control/Music/". You can
+                        optionally limit the depth: 0=root, 1=artist, 2=album,
+                        n=subfoldername, default=-1 (No Limit).
+  -i [ID3_TEMPLATE], --auto-id3-playlists [ID3_TEMPLATE]
+                        Generate automatic playlists based on the id3 tags of
+                        any music added to the iPod. You can optionally
+                        specify a template string based on which id3 tags are
+                        used to generate playlists. For eg. '{artist} -
+                        {album}' will use the pair of artist and album to
+                        group tracks under one playlist. Similarly '{genre}'
+                        will group tracks based on their genre tag. Default
+                        template used is '{artist}'
+  -v, --verbose         Show verbose output of database generation.
 ```
 
-#### Additions to the original
-* Option to disable voiceover
-* Initialize the IPod Directory tree
-* Using the --rename-unicode flag, filenames with strange characters and different language are renamed which avoids the script to crash with a Unicode Error
-
 #### Dependencies
 
 This script requires:
+* [Python 3](https://www.python.org/download/releases/3.0/)
 
-* [Python 2.7](http://www.python.org/download/releases/2.7/)
+Optional album/artist and auto-id3-playlists support:
 * [Mutagen](https://code.google.com/p/mutagen/)
-* [PicoSpeaker](http://picospeaker.tk/readme.php) -- for non-Russian files
-* [RHVoice (master branch, 3e31edced402a08771d2c48c73213982cbe9333e)](https://github.com/Olga-Yakovleva/RHVoice) -- for Russian files only
-* [SoX](http://sox.sourceforge.net) -- for Russian files only
+
+Optional Voiceover support
+* [eSpeak](http://espeak.sourceforge.net/)
+* [PicoSpeaker](http://picospeaker.tk/readme.php)
+* [RHVoice (master branch, 3e31edced402a08771d2c48c73213982cbe9333e)](https://github.com/Olga-Yakovleva/RHVoice) -- (Russian files only)
+* [SoX](http://sox.sourceforge.net) -- (Russian files)
+* say (macOS)
 
 ##### Ubuntu
 
-`apt-get install python-mutagen libttspico*`
+`apt-get install python3 python-mutagen libttspico*`
 
 ##### Arch Linux
 
-From the **Extra** repository: `pacman -S python2 mutagen` and from the AUR: `svox-pico-bin` ([link](https://aur.archlinux.org/packages/svox-pico-bin/))
+From the **Extra** repository: `pacman -S python` and optional `pacman -S python-mutagen espeak` or from the AUR: `svox-pico-bin` ([link](https://aur.archlinux.org/packages/svox-pico-bin/))
+
+You can also [install the script from AUR](https://aur.archlinux.org/packages/ipod-shuffle-4g/).
 
 ##### Gentoo Linux
 
 ```bash
-PYTHON_TARGETS="python2_7" emerge -av media-libs/mutagen
+PYTHON_TARGETS="python3" emerge -av media-libs/mutagen
 layman --add=ikelos
 layman --overlays="https://raw.githubusercontent.com/ahippo/rhvoice-gentoo-overlay/master/repositories.xml" --fetch --add=ahippo-rhvoice-overlay
 ACCEPT_KEYWORDS="~amd64" emerge -av app-accessibility/svox app-accessibility/rhvoice
 ```
 References to the overlays above: [ikelos](http://git.overlays.gentoo.org/gitweb/?p=dev/ikelos.git;a=summary), [ahippo-rhvoice-overlay](https://github.com/ahippo/rhvoice-gentoo-overlay)
 
-##Tips and Tricks
+## Tips and Tricks
 
 #### Disable trash for IPod
 To avoid that linux moves deleted files into trash you can create an empty file `.Trash-1000`.
 This forces linux to delete the files permanently instead of moving them to the trash.
 Of course you can also use `shift + delete` to permanently delete files without this trick.
+The file can be found in the [extras](extras) folder.
+
+#### Compress/Convert your music files
+([#11](https://github.com/nims11/IPod-Shuffle-4g/issues/11)) Shuffle is short on storage, and you might want to squeeze in more of your collection by sacrificing some bitrate off your files. In rarer cases, you might also possess music in formats not supported by your ipod. Although `ffmpeg` can handle almost all your needs, if you are looking for a friendly alternative, try [Soundconverter](http://soundconverter.org/).
 
 #### Use Rhythmbox to manage your music and playlists
 As described [in the blog post](https://nims11.wordpress.com/2013/10/12/ipod-shuffle-4g-under-linux/)
@@ -76,6 +104,7 @@ Simply place a file called `.is_audio_player` into the root directory of your IP
 name="Name's IPOD"
 audio_folders=iPod_Control/Music/
 ```
+The file can be found in the [extras](extras) folder.
 
 Now disable the IPod plugin of Rhythmbox and enable the MTP plugin instead.
 You can use Rythmbox now to generate playlists and sync them to your IPod.
@@ -91,6 +120,17 @@ In all cases you can try to update Rythmbox to the latest version, sync again or
 If you want to use this script on different computers it makes sense
 to simply copy the script into the IPod's root directory.
 
+#### Format/Restore/Recover IPod
+([#41](https://github.com/nims11/IPod-Shuffle-4g/issues/41)) If you formatted your IPod wrong and lost all data you can still recover it.
+It is important to **not use MBR/GPT**. You need to directly create a **Fat16 Filesystem**:
+
+```bash
+sudo mkfs.vfat -I -F 16 -n IPOD /dev/sdX
+```
+
+Run this script to generate the new database. All missing sound files should be regenrated by the IPod on next use.
+Your IPod should work and play music again now.
+
 ## TODO
 * Last.fm Scrobbler
 * Qt frontend

docs/iTunesSD3gen.md

@@ -41,9 +41,10 @@ Here's the general layout of an iTunesSD file:<br>
 </td>
         <td>4<br>
 </td>
-        <td>?<br>
+        <td>Version number?<br>
 </td>
-        <td><span style="font-family: 'Courier New',Courier,monospace;">0x03000002</span><br>
+        <td><span style="font-family: 'Courier New',Courier,monospace;">0x02000003<br>
+        Old values:<br>0x02010001<br>Gen 2:<br>0x010600<br>0x010800<br></span><br>
 </td>
         <td><span style="font-family: 'Courier New',Courier,monospace;">03 00 00 02</span><br>
 </td>
@@ -364,7 +365,7 @@ Here's the general layout of an iTunesSD file:<br>
 </td>
         <td>4<br>
 </td>
-        <td><br>
+        <td>Rythmbox IPod plugin sets this value always 0.<br>
 </td>
         <td><span style="font-family: 'Courier New',Courier,monospace;">112169</span><br>
 </td>

extras/.is_audio_player

@@ -0,0 +1,2 @@
+name="iPod Shuffle"
+audio_folders=iPod_Control/Music/

ipod-shuffle-4g.py

@@ -1,12 +1,11 @@
-#!/usr/bin/env python2.7
-# -*- coding: utf-8 -*-
+#!/usr/bin/env python3
+
+# Builtin libraries
 import sys
 import struct
-import urllib
+import urllib.request, urllib.parse, urllib.error
 import os
 import hashlib
-import mutagen
-import binascii
 import subprocess
 import collections
 import errno
@@ -16,6 +15,12 @@ import re
 import tempfile
 import signal
 
+# External libraries
+try:
+    import mutagen
+except ImportError:
+    mutagen = None
+
 audio_ext = (".mp3", ".m4a", ".m4b", ".m4p", ".aa", ".wav")
 list_ext = (".pls", ".m3u")
 def make_dir_if_absent(path):
@@ -27,19 +32,19 @@ def make_dir_if_absent(path):
 
 def raises_unicode_error(str):
     try:
-        str.decode('utf-8').encode('latin-1')
+        str.encode('latin-1')
         return False
     except (UnicodeEncodeError, UnicodeDecodeError):
         return True
 
 def hash_error_unicode(item):
-    return "".join(["{0:02X}".format(ord(x)) for x in reversed(hashlib.md5(item).digest()[:8])])
-    pass
+    item_bytes = item.encode('utf-8')
+    return "".join(["{0:02X}".format(ord(x)) for x in reversed(hashlib.md5(item_bytes).hexdigest()[:8])])
 
 def validate_unicode(path):
     path_list = path.split('/')
     last_raise = False
-    for i in xrange(len(path_list)):
+    for i in range(len(path_list)):
         if raises_unicode_error(path_list[i]):
             path_list[i] = hash_error_unicode(path_list[i])
             last_raise = True
@@ -51,29 +56,78 @@ def validate_unicode(path):
 def exec_exists_in_path(command):
     with open(os.devnull, 'w') as FNULL:
         try:
-            subprocess.call([command], stdout=FNULL, stderr=subprocess.STDOUT)
-            return True
+            with open(os.devnull, 'r') as RFNULL:
+                subprocess.call([command], stdout=FNULL, stderr=subprocess.STDOUT, stdin=RFNULL)
+                return True
         except OSError as e:
             return False
 
+def splitpath(path):
+    return path.split(os.sep)
+
+def get_relpath(path, basepath):
+    commonprefix = os.sep.join(os.path.commonprefix(list(map(splitpath, [path, basepath]))))
+    return os.path.relpath(path, commonprefix)
+
+def is_path_prefix(prefix, path):
+    return prefix == os.sep.join(os.path.commonprefix(list(map(splitpath, [prefix, path]))))
+
+def group_tracks_by_id3_template(tracks, template):
+    grouped_tracks_dict = {}
+    template_vars = set(re.findall(r'{.*?}', template))
+    for track in tracks:
+        try:
+            id3_dict = mutagen.File(track, easy=True)
+        except:
+            id3_dict = {}
+
+        key = template
+        single_var_present = False
+        for var in template_vars:
+            val = id3_dict.get(var[1:-1], [''])[0]
+            if len(val) > 0:
+                single_var_present = True
+            key = key.replace(var, val)
+
+        if single_var_present:
+            if key not in grouped_tracks_dict:
+                grouped_tracks_dict[key] = []
+            grouped_tracks_dict[key].append(track)
+
+    return sorted(grouped_tracks_dict.items())
+
 class Text2Speech(object):
-    valid_tts = {'pico2wave': True, 'RHVoice': True}
+    valid_tts = {'pico2wave': True, 'RHVoice': True, 'espeak': True, 'say': True}
 
     @staticmethod
     def check_support():
         voiceoverAvailable = False
 
+        # Check for macOS say voiceover
+        if not exec_exists_in_path("say"):
+            Text2Speech.valid_tts['say'] = False
+            print("Warning: macOS say not found, voicever won't be generated using it.")
+        else:
+            voiceoverAvailable = True
+
         # Check for pico2wave voiceover
         if not exec_exists_in_path("pico2wave"):
             Text2Speech.valid_tts['pico2wave'] = False
-            print "Error executing pico2wave, voicever won't be generated using it."
+            print("Warning: pico2wave not found, voicever won't be generated using it.")
+        else:
+            voiceoverAvailable = True
+
+        # Check for espeak voiceover
+        if not exec_exists_in_path("espeak"):
+            Text2Speech.valid_tts['espeak'] = False
+            print("Warning: espeak not found, voicever won't be generated using it.")
         else:
             voiceoverAvailable = True
 
         # Check for Russian RHVoice voiceover
         if not exec_exists_in_path("RHVoice"):
             Text2Speech.valid_tts['RHVoice'] = False
-            print "Warning: Error executing RHVoice, Russian voicever won't be generated."
+            print("Warning: RHVoice not found, Russian voicever won't be generated.")
         else:
             voiceoverAvailable = True
 
@@ -83,26 +137,33 @@ class Text2Speech(object):
 
     @staticmethod
     def text2speech(out_wav_path, text):
-        # Skip voiceover geneartion if a track with the same name is used.
+        # Skip voiceover generation if a track with the same name is used.
         # This might happen with "Track001" or "01. Intro" names for example.
         if os.path.isfile(out_wav_path):
-            print "Using existing", out_wav_path
+            verboseprint("Using existing", out_wav_path)
             return True
 
         # ensure we deal with unicode later
-        if not isinstance(text, unicode):
-            text = unicode(text, 'utf-8')
+        if not isinstance(text, str):
+            text = str(text, 'utf-8')
         lang = Text2Speech.guess_lang(text)
         if lang == "ru-RU":
             return Text2Speech.rhvoice(out_wav_path, text)
         else:
-            return Text2Speech.pico2wave(out_wav_path, text)
+            if Text2Speech.pico2wave(out_wav_path, text):
+                return True
+            elif Text2Speech.espeak(out_wav_path, text):
+                return True
+            elif Text2Speech.say(out_wav_path, text):
+                return True
+            else:
+                return False
 
     # guess-language seems like an overkill for now
     @staticmethod
     def guess_lang(unicodetext):
         lang = 'en-GB'
-        if re.search(u"[А-Яа-я]", unicodetext) is not None:
+        if re.search("[А-Яа-я]", unicodetext) is not None:
             lang = 'ru-RU'
         return lang
 
@@ -110,7 +171,21 @@ class Text2Speech(object):
     def pico2wave(out_wav_path, unicodetext):
         if not Text2Speech.valid_tts['pico2wave']:
             return False
-        subprocess.call(["pico2wave", "-l", "en-GB", "-w", out_wav_path, unicodetext])
+        subprocess.call(["pico2wave", "-l", "en-GB", "-w", out_wav_path, '--', unicodetext])
+        return True
+
+    @staticmethod
+    def say(out_wav_path, unicodetext):
+        if not Text2Speech.valid_tts['say']:
+            return False
+        subprocess.call(["say", "-o", out_wav_path, '--data-format=LEI16', '--file-format=WAVE', '--', unicodetext])
+        return True
+
+    @staticmethod
+    def espeak(out_wav_path, unicodetext):
+        if not Text2Speech.valid_tts['espeak']:
+            return False
+        subprocess.call(["espeak", "-v", "english_rp", "-s", "150", "-w", out_wav_path, '--', unicodetext])
         return True
 
     @staticmethod
@@ -136,12 +211,13 @@ class Record(object):
         self.parent = parent
         self._struct = collections.OrderedDict([])
         self._fields = {}
-        self.voiceover = parent.voiceover
+        self.track_voiceover = parent.track_voiceover
+        self.playlist_voiceover = parent.playlist_voiceover
         self.rename = parent.rename
         self.trackgain = parent.trackgain
 
     def __getitem__(self, item):
-        if item not in self._struct.keys():
+        if item not in list(self._struct.keys()):
             raise KeyError
         return self._fields.get(item, self._struct[item][1])
 
@@ -149,18 +225,16 @@ class Record(object):
         self._fields[item] = value
 
     def construct(self):
-        output = ""
-        for i in self._struct.keys():
+        output = bytes()
+        for i in list(self._struct.keys()):
             (fmt, default) = self._struct[i]
-            if fmt == "4s":
-                fmt, default = "I", int(binascii.hexlify(default), 16)
             output += struct.pack("<" + fmt, self._fields.get(i, default))
         return output
 
     def text_to_speech(self, text, dbid, playlist = False):
-        if self.voiceover:
+        if self.track_voiceover and not playlist or self.playlist_voiceover and playlist:
             # Create the voiceover wav file
-            fn = "".join(["{0:02X}".format(ord(x)) for x in reversed(dbid)])
+            fn = ''.join(format(x, '02x') for x in reversed(dbid))
             path = os.path.join(self.base, "iPod_Control", "Speakable", "Tracks" if not playlist else "Playlists", fn + ".wav")
             return Text2Speech.text2speech(path, text)
         return False
@@ -186,7 +260,7 @@ class Record(object):
 
     @property
     def base(self):
-        return self.shuffledb.base
+        return self.shuffledb.path
 
     @property
     def tracks(self):
@@ -210,19 +284,19 @@ class TunesSD(Record):
         self.track_header = TrackHeader(self)
         self.play_header = PlaylistHeader(self)
         self._struct = collections.OrderedDict([
-                           ("header_id", ("4s", "shdb")),
+                           ("header_id", ("4s", b"bdhs")), # shdb
                            ("unknown1", ("I", 0x02000003)),
                            ("total_length", ("I", 64)),
                            ("total_number_of_tracks", ("I", 0)),
                            ("total_number_of_playlists", ("I", 0)),
                            ("unknown2", ("Q", 0)),
                            ("max_volume", ("B", 0)),
-                           ("voiceover_enabled", ("B", int(self.voiceover))),
+                           ("voiceover_enabled", ("B", int(self.track_voiceover))),
                            ("unknown3", ("H", 0)),
                            ("total_tracks_without_podcasts", ("I", 0)),
                            ("track_header_offset", ("I", 64)),
                            ("playlist_header_offset", ("I", 0)),
-                           ("unknown4", ("20s", "\x00" * 20)),
+                           ("unknown4", ("20s", b"\x00" * 20)),
                                                ])
 
     def construct(self):
@@ -247,7 +321,7 @@ class TrackHeader(Record):
         self.base_offset = 0
         Record.__init__(self, parent)
         self._struct = collections.OrderedDict([
-                           ("header_id", ("4s", "shth")),
+                           ("header_id", ("4s", b"hths")), # shth
                            ("total_length", ("I", 0)),
                            ("number_of_tracks", ("I", 0)),
                            ("unknown1", ("Q", 0)),
@@ -259,10 +333,10 @@ class TrackHeader(Record):
         output = Record.construct(self)
 
         # Construct the underlying tracks
-        track_chunk = ""
+        track_chunk = bytes()
         for i in self.tracks:
             track = Track(self)
-            print "[*] Adding track", i
+            verboseprint("[*] Adding track", i)
             track.populate(i)
             output += struct.pack("I", self.base_offset + self["total_length"] + len(track_chunk))
             track_chunk += track.construct()
@@ -273,13 +347,13 @@ class Track(Record):
     def __init__(self, parent):
         Record.__init__(self, parent)
         self._struct = collections.OrderedDict([
-                           ("header_id", ("4s", "shtr")),
+                           ("header_id", ("4s", b"rths")), # shtr
                            ("header_length", ("I", 0x174)),
                            ("start_at_pos_ms", ("I", 0)),
                            ("stop_at_pos_ms", ("I", 0)),
                            ("volume_gain", ("I", int(self.trackgain))),
                            ("filetype", ("I", 1)),
-                           ("filename", ("256s", "\x00" * 256)),
+                           ("filename", ("256s", b"\x00" * 256)),
                            ("bookmark", ("I", 0)),
                            ("dontskip", ("B", 1)),
                            ("remember", ("B", 0)),
@@ -297,41 +371,49 @@ class Track(Record):
                            ("unknown4", ("Q", 0)),
                            ("dbid", ("8s", 0)),
                            ("artistid", ("I", 0)),
-                           ("unknown5", ("32s", "\x00" * 32)),
+                           ("unknown5", ("32s", b"\x00" * 32)),
                            ])
 
     def populate(self, filename):
-        self["filename"] = self.path_to_ipod(filename)
+        self["filename"] = self.path_to_ipod(filename).encode('utf-8')
 
         if os.path.splitext(filename)[1].lower() in (".m4a", ".m4b", ".m4p", ".aa"):
             self["filetype"] = 2
 
         text = os.path.splitext(os.path.basename(filename))[0]
-        audio = mutagen.File(filename, easy = True)
-        if audio:
-            self["stop_at_pos_ms"] = int(audio.info.length * 1000)
 
-            artist = audio.get("artist", [u"Unknown"])[0]
-            if artist in self.artists:
-                self["artistid"] = self.artists.index(artist)
-            else:
-                self["artistid"] = len(self.artists)
-                self.artists.append(artist)
+        # Try to get album and artist information with mutagen
+        if mutagen:
+            audio = None
+            try:
+                audio = mutagen.File(filename, easy = True)
+            except:
+                print("Error calling mutagen. Possible invalid filename/ID3Tags (hyphen in filename?)")
+            if audio:
+                # Note: Rythmbox IPod plugin sets this value always 0.
+                self["stop_at_pos_ms"] = int(audio.info.length * 1000)
 
-            album = audio.get("album", [u"Unknown"])[0]
-            if album in self.albums:
-                self["albumid"] = self.albums.index(album)
-            else:
-                self["albumid"] = len(self.albums)
-                self.albums.append(album)
+                artist = audio.get("artist", ["Unknown"])[0]
+                if artist in self.artists:
+                    self["artistid"] = self.artists.index(artist)
+                else:
+                    self["artistid"] = len(self.artists)
+                    self.artists.append(artist)
 
-            if audio.get("title", "") and audio.get("artist", ""):
-                text = u" - ".join(audio.get("title", u"") + audio.get("artist", u""))
+                album = audio.get("album", ["Unknown"])[0]
+                if album in self.albums:
+                    self["albumid"] = self.albums.index(album)
+                else:
+                    self["albumid"] = len(self.albums)
+                    self.albums.append(album)
+
+                if audio.get("title", "") and audio.get("artist", ""):
+                    text = " - ".join(audio.get("title", "") + audio.get("artist", ""))
 
         # Handle the VoiceOverData
-        if isinstance(text, unicode):
+        if isinstance(text, str):
             text = text.encode('utf-8', 'ignore')
-        self["dbid"] = hashlib.md5(text).digest()[:8] #pylint: disable-msg=E1101
+        self["dbid"] = hashlib.md5(text).digest()[:8]
         self.text_to_speech(text, self["dbid"])
 
 class PlaylistHeader(Record):
@@ -339,19 +421,19 @@ class PlaylistHeader(Record):
         self.base_offset = 0
         Record.__init__(self, parent)
         self._struct = collections.OrderedDict([
-                          ("header_id", ("4s", "shph")),
+                          ("header_id", ("4s", b"hphs")), #shph
                           ("total_length", ("I", 0)),
                           ("number_of_playlists", ("I", 0)),
-                          ("number_of_non_podcast_lists", ("2s", "\xFF\xFF")),
-                          ("number_of_master_lists", ("2s", "\x01\x00")),
-                          ("number_of_non_audiobook_lists", ("2s", "\xFF\xFF")),
-                          ("unknown2", ("2s", "\x00" * 2)),
+                          ("number_of_non_podcast_lists", ("2s", b"\xFF\xFF")),
+                          ("number_of_master_lists", ("2s", b"\x01\x00")),
+                          ("number_of_non_audiobook_lists", ("2s", b"\xFF\xFF")),
+                          ("unknown2", ("2s", b"\x00" * 2)),
                                               ])
 
-    def construct(self, tracks): #pylint: disable-msg=W0221
+    def construct(self, tracks):
         # Build the master list
         masterlist = Playlist(self)
-        print "[+] Adding master playlist"
+        verboseprint("[+] Adding master playlist")
         masterlist.set_master(tracks)
         chunks = [masterlist.construct(tracks)]
 
@@ -359,14 +441,14 @@ class PlaylistHeader(Record):
         playlistcount = 1
         for i in self.lists:
             playlist = Playlist(self)
-            print "[+] Adding playlist", i
+            verboseprint("[+] Adding playlist", (i[0] if type(i) == type(()) else i))
             playlist.populate(i)
             construction = playlist.construct(tracks)
             if playlist["number_of_songs"] > 0:
                 playlistcount += 1
                 chunks += [construction]
             else:
-                print "Error: Playlist does not contain a single track. Skipping playlist."
+                print("Error: Playlist does not contain a single track. Skipping playlist.")
 
         self["number_of_playlists"] = playlistcount
         self["total_length"] = 0x14 + (self["number_of_playlists"] * 4)
@@ -379,27 +461,27 @@ class PlaylistHeader(Record):
             output += struct.pack("I", offset)
             offset += len(chunks[i])
 
-        return output + "".join(chunks)
+        return output + b"".join(chunks)
 
 class Playlist(Record):
     def __init__(self, parent):
         self.listtracks = []
         Record.__init__(self, parent)
         self._struct = collections.OrderedDict([
-                          ("header_id", ("4s", "shpl")),
+                          ("header_id", ("4s", b"lphs")), # shpl
                           ("total_length", ("I", 0)),
                           ("number_of_songs", ("I", 0)),
                           ("number_of_nonaudio", ("I", 0)),
-                          ("dbid", ("8s", "\x00" * 8)),
+                          ("dbid", ("8s", b"\x00" * 8)),
                           ("listtype", ("I", 2)),
-                          ("unknown1", ("16s", "\x00" * 16))
+                          ("unknown1", ("16s", b"\x00" * 16))
                                               ])
 
     def set_master(self, tracks):
         # By default use "All Songs" builtin voiceover (dbid all zero)
         # Else generate alternative "All Songs" to fit the speaker voice of other playlists
-        if self.voiceover and Text2Speech.valid_tts['pico2wave']:
-            self["dbid"] = hashlib.md5("masterlist").digest()[:8] #pylint: disable-msg=E1101
+        if self.playlist_voiceover and (Text2Speech.valid_tts['pico2wave'] or Text2Speech.valid_tts['espeak'] or Text2Speech.valid_tts['say']):
+            self["dbid"] = hashlib.md5(b"masterlist").digest()[:8]
             self.text_to_speech("All songs", self["dbid"], True)
         self["listtype"] = 1
         self.listtracks = tracks
@@ -420,7 +502,7 @@ class Playlist(Record):
             dataarr = i.strip().split("=", 1)
             if dataarr[0].lower().startswith("file"):
                 num = int(dataarr[0][4:])
-                filename = urllib.unquote(dataarr[1]).strip()
+                filename = urllib.parse.unquote(dataarr[1]).strip()
                 if filename.lower().startswith('file://'):
                     filename = filename[7:]
                 if self.rename:
@@ -429,39 +511,70 @@ class Playlist(Record):
         listtracks = [ x for (_, x) in sorted(sorttracks) ]
         return listtracks
 
+    def populate_directory(self, playlistpath, recursive = True):
+        # Add all tracks inside the folder and its subfolders recursively.
+        # Folders containing no music and only a single Album
+        # would generate duplicated playlists. That is intended and "wont fix".
+        # Empty folders (inside the music path) will generate an error -> "wont fix".
+        listtracks = []
+        for (dirpath, dirnames, filenames) in os.walk(playlistpath):
+            dirnames.sort()
+
+            # Ignore any hidden directories
+            if "/." not in dirpath:
+                for filename in sorted(filenames, key = lambda x: x.lower()):
+                    # Only add valid music files to playlist
+                    if os.path.splitext(filename)[1].lower() in (".mp3", ".m4a", ".m4b", ".m4p", ".aa", ".wav"):
+                        fullPath = os.path.abspath(os.path.join(dirpath, filename))
+                        listtracks.append(fullPath)
+            if not recursive:
+                break
+        return listtracks
+
     def remove_relatives(self, relative, filename):
         base = os.path.dirname(os.path.abspath(filename))
         if not os.path.exists(relative):
             relative = os.path.join(base, relative)
         fullPath = relative
-        ipodpath = self.parent.parent.parent.path
-        relPath = fullPath[fullPath.index(ipodpath)+len(ipodpath)+1:].lower()
-        fullPath = os.path.abspath(os.path.join(ipodpath, relPath))
         return fullPath
 
-    def populate(self, filename):
-        with open(filename, 'rb') as f:
-            data = f.readlines()
+    def populate(self, obj):
+        # Create a playlist of the folder and all subfolders
+        if type(obj) == type(()):
+            self.listtracks = obj[1]
+            text = obj[0]
+        else:
+            filename = obj
+            if os.path.isdir(filename):
+                self.listtracks = self.populate_directory(filename)
+                text = os.path.splitext(os.path.basename(filename))[0]
+            else:
+                # Read the playlist file
+                with open(filename, 'r', errors="replace") as f:
+                    data = f.readlines()
 
-        extension = os.path.splitext(filename)[1].lower()
-        if extension == '.pls':
-            self.listtracks = self.populate_pls(data)
-        elif extension == '.m3u':
-            self.listtracks = self.populate_m3u(data)
-        # Ensure all paths are not relative to the playlist file
-        for i in range(len(self.listtracks)):
-            self.listtracks[i] = self.remove_relatives(self.listtracks[i], filename)
+                extension = os.path.splitext(filename)[1].lower()
+                if extension == '.pls':
+                    self.listtracks = self.populate_pls(data)
+                elif extension == '.m3u':
+                    self.listtracks = self.populate_m3u(data)
+                else:
+                    raise
+
+                # Ensure all paths are not relative to the playlist file
+                for i in range(len(self.listtracks)):
+                    self.listtracks[i] = self.remove_relatives(self.listtracks[i], filename)
+                text = os.path.splitext(os.path.basename(filename))[0]
 
         # Handle the VoiceOverData
-        text = os.path.splitext(os.path.basename(filename))[0]
-        self["dbid"] = hashlib.md5(text).digest()[:8] #pylint: disable-msg=E1101
+        self["dbid"] = hashlib.md5(text.encode('utf-8')).digest()[:8]
         self.text_to_speech(text, self["dbid"], True)
 
-    def construct(self, tracks): #pylint: disable-msg=W0221
+    def construct(self, tracks):
         self["total_length"] = 44 + (4 * len(self.listtracks))
         self["number_of_songs"] = 0
 
-        chunks = ""
+        chunks = bytes()
         for i in self.listtracks:
             path = self.ipod_to_path(i)
             position = -1
@@ -470,8 +583,8 @@ class Playlist(Record):
             except:
                 # Print an error if no track was found.
                 # Empty playlists are handeled in the PlaylistHeader class.
-                print "Error: Could not find track \"" + path + "\"."
-                print "Maybe its an invalid FAT filesystem name. Please fix your playlist. Skipping track."
+                print("Error: Could not find track \"" + path + "\".")
+                print("Maybe its an invalid FAT filesystem name. Please fix your playlist. Skipping track.")
             if position > -1:
                 chunks += struct.pack("I", position)
                 self["number_of_songs"] += 1
@@ -481,16 +594,19 @@ class Playlist(Record):
         return output + chunks
 
 class Shuffler(object):
-    def __init__(self, path, voiceover=True, rename=False, trackgain=0):
-        self.path, self.base = self.determine_base(path)
+    def __init__(self, path, track_voiceover=False, playlist_voiceover=False, rename=False, trackgain=0, auto_dir_playlists=None, auto_id3_playlists=None):
+        self.path = os.path.abspath(path)
         self.tracks = []
         self.albums = []
         self.artists = []
         self.lists = []
         self.tunessd = None
-        self.voiceover = voiceover
+        self.track_voiceover = track_voiceover
+        self.playlist_voiceover = playlist_voiceover
         self.rename = rename
         self.trackgain = trackgain
+        self.auto_dir_playlists = auto_dir_playlists
+        self.auto_id3_playlists = auto_id3_playlists
 
     def initialize(self):
       # remove existing voiceover files (they are either useless or will be overwritten anyway)
@@ -500,36 +616,59 @@ class Shuffler(object):
           make_dir_if_absent(os.path.join(self.path, dirname))
 
     def dump_state(self):
-        print "Shuffle DB state"
-        print "Tracks", self.tracks
-        print "Albums", self.albums
-        print "Artists", self.artists
-        print "Playlists", self.lists
-
-    def determine_base(self, path):
-        base = os.path.abspath(path)
-        # while not os.path.ismount(base):
-        #     base = os.path.dirname(base)
-        return base, base
+        print("Shuffle DB state")
+        print("Tracks", self.tracks)
+        print("Albums", self.albums)
+        print("Artists", self.artists)
+        print("Playlists", self.lists)
 
     def populate(self):
         self.tunessd = TunesSD(self)
         for (dirpath, dirnames, filenames) in os.walk(self.path):
             dirnames.sort()
+            relpath = get_relpath(dirpath, self.path)
             # Ignore the speakable directory and any hidden directories
-            if "ipod_control/speakable" not in dirpath.lower() and "/." not in dirpath.lower():
+            if not is_path_prefix("iPod_Control/Speakable", relpath) and "/." not in dirpath:
                 for filename in sorted(filenames, key = lambda x: x.lower()):
-                    fullPath = os.path.abspath(os.path.join(dirpath, filename))
-                    relPath = fullPath[fullPath.index(self.path)+len(self.path)+1:].lower()
-                    fullPath = os.path.abspath(os.path.join(self.path, relPath));
-                    if os.path.splitext(filename)[1].lower() in (".mp3", ".m4a", ".m4b", ".m4p", ".aa", ".wav"):
-                        self.tracks.append(fullPath)
-                    if os.path.splitext(filename)[1].lower() in (".pls", ".m3u"):
-                        self.lists.append(os.path.abspath(os.path.join(dirpath, filename)))
+                    # Ignore hidden files
+                    if not filename.startswith("."):
+                        fullPath = os.path.abspath(os.path.join(dirpath, filename))
+                        if os.path.splitext(filename)[1].lower() in (".mp3", ".m4a", ".m4b", ".m4p", ".aa", ".wav"):
+                            self.tracks.append(fullPath)
+                        if os.path.splitext(filename)[1].lower() in (".pls", ".m3u"):
+                            self.lists.append(fullPath)
+
+            # Create automatic playlists in music directory.
+            # Ignore the (music) root and any hidden directories.
+            if self.auto_dir_playlists and "iPod_Control/Music/" in dirpath and "/." not in dirpath:
+                # Only go to a specific depth. -1 is unlimted, 0 is ignored as there is already a master playlist.
+                depth = dirpath[len(self.path) + len(os.path.sep):].count(os.path.sep) - 1
+                if self.auto_dir_playlists < 0 or depth <= self.auto_dir_playlists:
+                    self.lists.append(os.path.abspath(dirpath))
+
+        if self.auto_id3_playlists != None:
+            if mutagen:
+                for grouped_list in group_tracks_by_id3_template(self.tracks, self.auto_id3_playlists):
+                    self.lists.append(grouped_list)
+            else:
+                print("Error: No mutagen found. Cannot generate auto-id3-playlists.")
+                sys.exit(1)
 
     def write_database(self):
-        with open(os.path.join(self.base, "iPod_Control", "iTunes", "iTunesSD"), "wb") as f:
-            f.write(self.tunessd.construct())
+        print("Writing database. This may take a while...")
+        with open(os.path.join(self.path, "iPod_Control", "iTunes", "iTunesSD"), "wb") as f:
+            try:
+                f.write(self.tunessd.construct())
+            except IOError as e:
+                print("I/O error({0}): {1}".format(e.errno, e.strerror))
+                print("Error: Writing iPod database failed.")
+                sys.exit(1)
+
+        print("Database written successfully:")
+        print("Tracks", len(self.tracks))
+        print("Albums", len(self.albums))
+        print("Artists", len(self.artists))
+        print("Playlists", len(self.lists))
 
 #
 # Read all files from the directory
@@ -548,7 +687,7 @@ def check_unicode(path):
                 if raises_unicode_error(item):
                     src = os.path.join(path, item)
                     dest = os.path.join(path, hash_error_unicode(item)) + os.path.splitext(item)[1].lower()
-                    print 'Renaming %s -> %s' % (src, dest)
+                    print('Renaming %s -> %s' % (src, dest))
                     os.rename(src, dest)
         else:
             ret_flag = (check_unicode(os.path.join(path, item)) or ret_flag)
@@ -556,7 +695,7 @@ def check_unicode(path):
                 src = os.path.join(path, item)
                 new_name = hash_error_unicode(item)
                 dest = os.path.join(path, new_name)
-                print 'Renaming %s -> %s' % (src, dest)
+                print('Renaming %s -> %s' % (src, dest))
                 os.rename(src, dest)
     return ret_flag
 
@@ -572,36 +711,86 @@ def nonnegative_int(string):
 
 def checkPathValidity(path):
     if not os.path.isdir(result.path):
-        print "Error finding IPod directory. Maybe it is not connected or mounted?"
+        print("Error finding IPod directory. Maybe it is not connected or mounted?")
         sys.exit(1)
 
     if not os.access(result.path, os.W_OK):
-        print 'Unable to get write permissions in the IPod directory'
+        print('Unable to get write permissions in the IPod directory')
         sys.exit(1)
 
 def handle_interrupt(signal, frame):
-    print "Interrupt detected, exiting..."
+    print("Interrupt detected, exiting...")
     sys.exit(1)
 
 if __name__ == '__main__':
     signal.signal(signal.SIGINT, handle_interrupt)
-    parser = argparse.ArgumentParser()
-    parser.add_argument('--disable-voiceover', action='store_true', help='Disable voiceover feature')
-    parser.add_argument('--rename-unicode', action='store_true', help='Rename files causing unicode errors, will do minimal required renaming')
-    parser.add_argument('--track-gain', type=nonnegative_int, default=0, help='Specify volume gain (0-99) for all tracks; 0 (default) means no gain and is usually fine; e.g. 60 is very loud even on minimal player volume')
+
+    parser = argparse.ArgumentParser(description=
+    'Python script for building the Track and Playlist database '
+    'for the newer gen IPod Shuffle. Version 1.5')
+
+    parser.add_argument('-t', '--track-voiceover', action='store_true',
+    help='Enable track voiceover feature')
+
+    parser.add_argument('-p', '--playlist-voiceover', action='store_true',
+    help='Enable playlist voiceover feature')
+
+    parser.add_argument('-u', '--rename-unicode', action='store_true',
+    help='Rename files causing unicode errors, will do minimal required renaming')
+
+    parser.add_argument('-g', '--track-gain', type=nonnegative_int, default='0',
+    help='Specify volume gain (0-99) for all tracks; '
+    '0 (default) means no gain and is usually fine; '
+    'e.g. 60 is very loud even on minimal player volume')
+
+    parser.add_argument('-d', '--auto-dir-playlists', type=int, default=None, const=-1, nargs='?',
+    help='Generate automatic playlists for each folder recursively inside '
+    '"IPod_Control/Music/". You can optionally limit the depth: '
+    '0=root, 1=artist, 2=album, n=subfoldername, default=-1 (No Limit).')
+
+    parser.add_argument('-i', '--auto-id3-playlists', type=str, default=None, metavar='ID3_TEMPLATE', const='{artist}', nargs='?',
+    help='Generate automatic playlists based on the id3 tags of any music '
+    'added to the iPod. You can optionally specify a template string '
+    'based on which id3 tags are used to generate playlists. For eg. '
+    '\'{artist} - {album}\' will use the pair of artist and album to group '
+    'tracks under one playlist. Similarly \'{genre}\' will group tracks based '
+    'on their genre tag. Default template used is \'{artist}\'')
+
+    parser.add_argument('-v', '--verbose', action='store_true',
+    help='Show verbose output of database generation.')
+
     parser.add_argument('path', help='Path to the IPod\'s root directory')
+
     result = parser.parse_args()
 
+    # Enable verbose printing if desired
+    verboseprint = print if result.verbose else lambda *a, **k: None
+
     checkPathValidity(result.path)
 
     if result.rename_unicode:
         check_unicode(result.path)
 
-    if not result.disable_voiceover and not Text2Speech.check_support():
-            print "Error: Did not find any voiceover program. Voiceover disabled."
-            result.disable_voiceover = True
+    if not mutagen:
+        print("Warning: No mutagen found. Database will not contain any album nor artist information.")
 
-    shuffle = Shuffler(result.path, voiceover=not result.disable_voiceover, rename=result.rename_unicode, trackgain=result.track_gain)
+    verboseprint("Playlist voiceover requested:", result.playlist_voiceover)
+    verboseprint("Track voiceover requested:", result.track_voiceover)
+    if (result.track_voiceover or result.playlist_voiceover):
+        if not Text2Speech.check_support():
+            print("Error: Did not find any voiceover program. Voiceover disabled.")
+            result.track_voiceover = False
+            result.playlist_voiceover = False
+        else:
+            verboseprint("Voiceover available.")
+
+    shuffle = Shuffler(result.path,
+                       track_voiceover=result.track_voiceover,
+                       playlist_voiceover=result.playlist_voiceover,
+                       rename=result.rename_unicode,
+                       trackgain=result.track_gain,
+                       auto_dir_playlists=result.auto_dir_playlists,
+                       auto_id3_playlists=result.auto_id3_playlists)
     shuffle.initialize()
     shuffle.populate()
     shuffle.write_database()